Research On MTPSA Control Method For Permanent Magnet Synchronous Motor Based On Improved Gradient Descent Method

In order to reduce the copper consumption of permanent magnet synchronous motor(PMSM)and improve its energy utilization rate,maximum torque per current ratio(MTPA)control is often used.For interior permanent magnet synchronous motor(IPMSM),a maximum torque per ampere(MTPSA)control method equivalent to MTPA control is proposed to obtain the minimum current amplitude under a given torque.In this dissertation,the mathematical model of IPMSM with traditional fixed parameters is improved,and the mathematical model of IPMSM considering parameter changes is obtained.The gradient descent algorithm is used to search the current vector angle satisfying MTPSA control online.Aiming at the problem that the search accuracy and algorithm execution time depend on the change of load torque in the MTPSA control process based on the traditional gradient descent method,the gradient descent method is improved.Firstly.the traditional MTPA control method is deeply studied,and its control effect is directly related to the main parameters of IPMSM(such as quadrature-direct axis inductance.permanent magnet flux linkage.etc.).However,the mathematical model of IPMSM commonly used in most cases does not consider the influence of parameter changes.Aiming at this problem,this dissertation first establishes a mathematical model of IPMSM considering the parameter changes caused by magnetic circuit saturation effect and cross-coupling effect.Using this model for simulation can get closer to the actual operation of the motor.Secondly,the MTPA and MTPSA control methods are deeply studied.Through the derivation and analysis of the two methods,it can be seen that MTPSA control and MTPA control are equivalent control.Compared with MTPA control,MTPSA control method does not need accurate stator resistance parameter R,and the objective function does not contain the main parameters of IPMSM.The required calculation parameters can be obtained by real-time sampling.Thirdly,through the analysis of the simulation results of MTPSA control based on gradient descent method,the search results and algorithm execution time are affected by the load torque.When the load torque is large,the convergence rate slows down,the motor is in a deep saturation state,the search result error is large,and the jitter amplitude is large near the optimal value.Aiming at this problem,this dissertation proposes an improved gradient descent method to realize MTPSA control.The variable gradient ’Q’ is introduced to describe the information of the current point towards the optimal point iteration,which can meet the requirement that the convergence rate is faster when the current point is far away from the optimal point to reduce the duration,and the convergence rate is slower when it is close to the optimal point to improve the search accuracy.At the same time,the convergence rate does not change with the change of load torque.Compared with the unimproved gradient descent method,it has a faster convergence rate and higher search accuracy.An IPMSM simulation model considering magnetic saturation and cross-coupling effect is built in MATLAB/Simulink software,and the stability and reliability of the model are verified.At the same time,MTPA control system simulation model based on traditional formula method,MTPSA control system simulation model based on gradient descent method and MTPSA control system simulation model based on improved gradient descent method are built,and the simulation results are compared and analyzed.Compared with MTPSA control based on gradient descent method,the MTPSA control method based on improved gradient descent method proposed in this dissertation reduces the dependence of search accuracy and algorithm execution time on load torque,which makes the search accuracy higher and the algorithm execution time shorter.Compared with the MTPA control based on the traditional formula method,since each iteration of the algorithm is iterated towards the optimal value under the current working condition,the proposed method does not depend on the motor parameters,and the accuracy of MTPA control is higher.